Wrong Or Merely Prohibited: Special Treatment of Strict Liability in Intuitive Moral Judgment

Wrong or Merely Prohibited: Special Treatment of Strict Liability in Intuitive Moral Judgment

Carly Giffin and Tania Lombrozo University of California, Berkeley

Most crimes in America require that the defendant have mens rea, Latin for “guilty mind.” However, mens rea is not legally required for strict liability crimes, such as speeding, for which someone is guilty even if ignorant or deceived about her speed. In 3 experiments involving participants responding to descriptive vignettes, we investigated whether the division of strict liability crimes in the law reflects an aspect of laypeople’s intuitive moral cognition. Experiment 1 (N ϭ 396; 236 male, 159 female, 1 other;

Mage ϭ 30) found evidence that it does: ignorance and deception were less mitigating for strict liability crimes than for “mens rea” crimes. Experiments 2 (N ϭ 413; 257 male, 154 female, 2 other; Mage ϭ 31) and 3 (N ϭ 404; 183 male, 221 female, Mage ϭ 35) revealed that strict liability crimes are not treated as pure moral violations, but additionally as violations of convention. We found that for strict liability crimes, ratings of moral wrongness and punishment were influenced to a greater extent by the fact that a rule had been violated, even when harm was kept constant, mirroring the legal distinction of malum prohibitum (wrong as prohibited) versus malum in se (wrong in itself). Further, we found that rules prohibiting strict liability crimes were judged more arbitrary than corresponding rules for “mens rea” crimes, and that this judgment was related to the role of mental states. Jointly, the findings suggest a surprising correspondence between the law and laypeople’s intuitive judgments.

Keywords: rule violation, knowledge, decision making, explanations, law

Supplemental materials: http://dx.doi.org/10.1037/lhb0000212.supp

“The first requirement of a sound body of law,” wrote Justice concept is referred to as “mens rea,” Latin for “guilty mind,” and Oliver Wendall Holmes, “is that it should correspond with the has been a part of criminal law for centuries (Blackstone, 1769). actual feelings and demands of the community, whether right or Psychological research confirms that, in most cases, a person’s wrong” (Holmes, 1881, p. 41). Whether one believes that the law intent is seen as a crucial variable in assigning moral responsibility should conform to the feelings of the community, or vice versa, the and punishment (e.g., Knobe, 2005; Malle & Knobe, 1997). For idea of a correspondence between the legal system and laypeople’s instance, putting white powder in someone’s coffee is judged intuitive judgments continues to have popular and academic sup- considerably more harshly when the powder was known to be port (e.g., Alicke, 2008). To the extent this correspondence holds, poison than when it was mistakenly thought to be sugar (Young & the psychology of moral judgment has important implications for Saxe, 2011). Thus, law and psychology agree about the important the law, and the law provides a rich source of information about role of mental states or mens rea, even if legal and intuitive human moral psychology (Hart & Honore, 1962; Mikhail, 2009). concepts are not perfectly aligned (Malle & Nelson, 2003). One important correspondence between the law and intuitive Notably, the law makes an exception concerning the role of judgments can be seen in the legal system’s requirements regard- mens rea for a specific class of crimes known as “strict liability” ing intent. Characteristically, a person cannot be found guilty of a crimes. In criminal law, strict liability crimes are crimes for which crime in America unless that person had some intent to commit the the prosecution does not have to prove that the defendant had the This document is copyrighted by the American Psychological Association or one of its alliedcrime publishers. or acted in a way that was negligent or reckless. Legally, this requisite mens rea with respect to at least one element of a crime, This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. and, in many cases, the defendant is not allowed to present evidence that she made a reasonable mistake or lacked relevant This article was published Online First September 15, 2016. knowledge or intent (Levenson, 1992). For instance, a person who Carly Giffin and Tania Lombrozo, Department of Psychology, Univer- is speeding can be found guilty without the prosecution proving sity of California, Berkeley. that she intended to speed or knew that she was speeding. We thank Elliot Turiel and the Concepts and Cognition Lab at the A variety of justifications for the imposition of strict liability University of California, Berkeley for helpful comments, as well as Bobbie have been offered. For example, some have argued that imposing Spellman and Elizabeth Gilbert for suggesting additional, relevant re- strict liability increases the incentive for the actor to use care or to search. This work was partially supported a James S. McDonnell Founda- consider whether the activity should be engaged in at all (Man- tion Scholar Award in Understanding Human Cognition to Tania Lom- brozo. chester, 1977; Singer, 1989), or that strict liability appropriately Correspondence concerning this article should be addressed to Carly Giffin, apportions risk on the only person who could have benefited from Department of Psychology, University of California, Berkeley, 3210 Tolman the action (Keating, 2006). These positions are often tied to dif- Hall, #1650, Berkeley, CA 94720. E-mail: [email protected] ferent justifications for punishment, such as the idea that punish-

707 708 GIFFIN AND LOMBROZO

ment should deter the behavior, rather than being imposed in violations than for “purity” violations, and many—but not all— retribution (Craswell & Calfee, 1986; Stahlkopf, Males, & Ma- offenses that have been historically treated as strict liability argu- callair, 2010). ably involve some element of purity (e.g., statutory rape, incest, Another explanation for strict liability comes from the Supreme bigamy, food adulteration). We therefore consider both public Court case of United States v. Morissette (1952), in which Justice welfare strict liability offenses that involve potential harm (speed- Jackson noted that, historically, many strict liability crimes were ing, selling drugs to minors) as well as moral-purity strict liability primarily malum prohibitum (wrong because they were prohibited) offenses that violate purity (statutory rape, incest). We compare rather than malum in se (wrong or evil in themselves). A crime that judgments concerning these crimes against those for closely is wrong merely by law, it was argued, does not require mens rea matched mens rea crimes that we likewise classify as public because the violation of the law, regardless of intent, is part of the welfare (reckless driving, drug distribution) or as moral-purity harm to be protected against. We will revisit this idea in Experi- (seducing a minor, first cousin marriage). ments 2 and 3. To investigate the role of mens rea in judging perpetrators of In the present work, we investigate whether the legal distinction these crimes, we presented participants with vignettes involving a between strict liability crimes and mens rea crimes reflects a transgression and varied whether the actor knew the fact that psychological distinction as well. One possibility is that strict rendered an action illegal—for example, that he was driving over liability crimes represent a departure from a general correspon- the speed limit, or that a sexual partner was underage. When the dence between the law and intuitive judgments. Indeed, many actor was ignorant, we additionally varied whether the ignorance scholars point out how unintuitive the imposition of strict liability resulted from inadvertent bad information or from intentional can seem (e.g., Mens Rea, 2013; Stanton-Ife, 2007). But a second deception; based on prior work, we expected that deception would possibility is that the distinction between strict liability and mens be more mitigating (Murray & Lombrozo, 2016; Phillips & Shaw, rea crimes reflects a psychological distinction that has not been 2015). Finally, we also included a condition in which mental states adequately recognized within the psychological literature. Based were unspecified. on our findings, this is what we will ultimately suggest. One primary hypothesis (Hypothesis 1) was that judgments In Experiment 1, our aim is to establish whether judgments concerning moral censure and punishment would depend on the concerning strict liability crimes are—on average—influenced actor’s mental state, with the harshest judgments when the actor less by the actor’s knowledge and intent than are mens rea crimes, transgressed knowingly. We also anticipated that being de- thus mirroring the law. Importantly, at no point are participants ceived would be more mitigating than mere ignorance (Hypoth- told or given any indication of which crimes are strict liability, nor esis 1a) and that these effects would be found for all crimes do they come to the experiments with this knowledge (see online (Hypothesis 1b). Supplemental Materials). Nonetheless, Experiment 1 reveals a Our secondary hypotheses concerned possible reasons why surprising degree of correspondence between the law’s special mens rea and strict liability crimes could receive different treat- treatment of strict liability crimes and laypeople’s intuitive moral ment. We tested the following predictions: (Hypothesis 2) that the judgments. mental state manipulation would influence judgments concerning In Experiments 1–3, we additionally explore potential bases for public welfare crimes more strongly than those concerning moral- differentiation between strict liability and mens rea crimes. Spe- purity crimes, (Hypothesis 3) that deterrence would be regarded as cifically, we investigate whether views about punishment differ for a more central reason for punishing strict liability crimes than for strict liability and mens rea crimes, including the roles of retribu- punishing mens rea crimes, and (Hypothesis 4) that when an tion (Experiment 1 and S1 in the Supplemental Materials; contact agent’s mental states were not specified, participants would more authors for full report), deterrence (Experiments 1 and S1), and often infer that the transgressor knew the key fact for strict liability incapacitation (Experiment S2; contact authors for full report). We crimes than for mens rea crimes. also investigate whether strict liability and mens rea crimes tend to license different inferences about the actor’s mental states (Exper- Method iments 1, 2, and S1), and whether relative to mens rea crimes, strict liability crimes are to a greater extent malum prohibitum (Exper- Participants. Three hundred ninety-six adults (236 male, 159

iments 2 and 3). We find support for only one of these proposals: female, one other/prefer not to specify, Mage ϭ 30, SD ϭ 10) This document is copyrighted by the American Psychological Association or one of its allied publishers. that relative to mens rea crimes, strict liability crimes are treated as participated in the study through Amazon Mechanical Turk. An This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. more malum prohibitum, and that this is because they tend to have additional 140 participants were tested, but were excluded for more arbitrary elements. failing catch questions (92) and due to a data collection error (52). In all studies, participation was restricted to workers with Inter- Experiment 1 net protocol addresses in the United States and with an approval rating of 95% or higher on previous Mechanical Turk tasks. Only Experiment 1 investigated whether laypeople judge strict liabil- U.S. citizens above the age of 18 were allowed to participate, to ity and mens rea crimes differently. Research in moral psychology better mimic the composition of an actual jury. Participants were suggests that mens rea is critical to moral judgment (e.g., Inbar, given $0.30 for their participation. Pizarro, & Cushman, 2012; Rai & Fiske, 2011), but there is also Materials and procedure. Participants were randomly as- evidence that a person’s knowledge and intentions are not equally signed to one of 32 distinct vignettes, about which they answered influential in evaluating all types of transgressions (Barrett et al., several questions (detailed below). Eight different crimes were 2016; Graham et al., 2011; Young & Saxe, 2011). Specifically, chosen, four of which are strict liability crimes in a majority of knowledge and intentions may be more important for “harm” states (statutory rape, incest, speeding, and selling drugs to minors) WRONG OR MERELY PROHIBITED 709

and four of which are mens rea crimes (seducing a minor, first cousin After reading the assigned vignette, it was removed and partic- marriage, reckless driving, and drug distribution; see Table 1). We ipants answered questions. First, participants rated our variables of chose mens rea crimes that were as similar as possible to the interest, moral censure and punishment, in a randomized order: corresponding strict liability crimes to make it more likely that differences, if found, would not reflect idiosyncratic properties of Moral censure: How morally wrong were [Actor’s] actions? the selected examples. Four of the crimes involved moral purity Participants indicated their answer on a scale from 1 (not at all (statutory rape, seducing a minor, incest, first cousin marriage) and morally wrong) to 7 (very morally wrong). four involved public welfare (speeding, reckless driving, selling drugs to minors, drug distribution). Punishment: How much punishment does [Actor] deserve? Each of the eight crimes was presented to a participant in one of four versions, which varied the principal actor’s mental states. In Participants indicated their answer on a scale from 1 (none at all) the knowing condition, participants were told that the actor knew to 7 (very much). We did not ask participants to evaluate guilt, as the relevant fact that made a crime illegal (e.g., the age of an in each vignette it was stipulated. underage sexual partner). In the unspecified condition, participants Second, to ensure that strict liability and mens rea vignettes were not given any information about the actor’s knowledge. In the were well matched along other dimensions that could impact unknowing condition, participants were told that the actor received judgments, participants rated the following questions on a separate false but well-intentioned information about the relevant fact (e.g., screen, in randomized order, from 1 (not at all/none) to 7 (very/a in the statutory rape vignette, another person honestly believed the great deal): underage victim to be an 18-year-old college student and said so). Disgust: How disgusting did you find [Actor’s] actions? In the deceived condition, the actor was deliberately deceived by another person (e.g., the victim lied about her age). Anger: How angry did [Actor’s] actions make you? Examples of the different mental state conditions for speeding are excerpted below (see online supplement for all materials): Harm: How much harm did [Actor’s] actions cause? Knowing: . . . Alan knew the speed limit was 55 miles per hour Third, to test whether differences in the evaluation of strict and that he was driving over the speed limit, but Alan didn’t liability and mens rea crimes stem from commitments about why slow down as the speedometer crept up to 70 miles per hour. punishment is appropriate for each crime type (Hypothesis 3), participants rated the statements below in a randomized order, Unspecified: Alan was driving along a two lane highway... from 1 (completely agree) to 7 (completely disagree): A police officer was parked alongside the highway with a speed gun pointed towards oncoming traffic . . . the officer... Deterrence: Punishing people who [commit this crime] is noticed that Alan was going 70 miles per hour. critically important to preventing acts of [this crime] in society in general. Unknowing: . . . Alan knew the speed limit was 55 miles per hour and was careful to check the speedometer periodically to Retribution: Even if it won’t prevent future [instances of this make sure the car wasn’t going faster than the speed limit... crime], a person who [commits this crime] deserves to be Alan had never had any reason to suspect the speedometer was punished for this action. broken but later found out that it was. Fourth, to ensure that participants were attending to the task and Deceived: . . . Alan knew the speed limit was 55 miles per reading materials carefully, participants answered two “catch ques- hour and was careful to check the speedometer periodically to tions,” easy true/false questions that assessed having read the make sure the car wasn’t going faster than the speed limit... vignette. Participants who answered either question incorrectly Alan was surprised to hear [he had been speeding]...Alan were excluded from further analyses. This screen included a third had never had any reason to suspect the speedometer was question about the actor’s knowledge (e.g., whether Alan knew he broken but later found out that his mechanic had intentionally was driving over the speed limit), which served as a mere com- tampered with it after Alan disputed a bill. prehension question for participants in the knowing, unknowing, This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Table 1 Crimes Used in Experiment 1

Moral classification Strict liability crimes Mens rea crimes

Moral Statutory rape (age of sexual partner) Seducing a minor (age of person receiving material) Incest (relatedness of sexual partner) First cousin marriage (relatedness of spouse) Public welfare Speeding (vehicle speed) Reckless driving (speed and attendant circumstances) Drugs to minors (age of customer) Drug distribution (nature of substance) Note. For each crime, participants were assigned to one of four mental state levels: knowing, unspecified, unknowing, or deceived. Each crime is followed in parentheses by the relevant knowledge that was varied across the mental state manipulation. 710 GIFFIN AND LOMBROZO

and deceived conditions, where knowledge or ignorance of the cantly higher than for the unknowing and deceived conditions vital fact—speed of the car, and so forth—was stipulated. It also (ps Ͻ .09, ds Ͼ .38), but that ratings for the unknowing and allowed us to test for differences in baseline inferences across deceived conditions were not significantly different from each crime types for the unspecified condition (Hypothesis 4). other for either dependent variable (ps Ͼ .30, ds Ͻ .15), challeng- Finally, participants justified their reaction to the story, received ing Hypothesis 1a. This suggests that misinformation had a miti- another attention check based on Oppenheimer, Meyvis, and gating effect, but that the source of the misinformation was not Davydenko (2009), and answered demographic questions. (Neither relevant. the justifications nor the demographic information was analyzed We also found significant effects of legal category for both for this article.) moral censure and punishment, with higher ratings for strict lia-

bility crimes (Mc ϭ 3.87, SDc ϭ 2.20; Mp ϭ 3.38, SDp ϭ 1.81) Results than for mens rea crimes (Mc ϭ 3.44, SDc ϭ 2.16; Mp ϭ 2.89, SDp ϭ 1.84; see Table 2). We hesitate to draw strong conclusions Disgust, anger, and harm. To ensure that our strict liability from these differences, as they could simply reflect our choice of and mens rea vignettes did not differ systematically in judged harm stimulus materials rather than the classes of crimes. Moreover, our or elicited anger and disgust, we performed a series of one-way primary predictions concern an interaction between mental state analyses of variance (ANOVAs) with legal category (strict liabil- and legal category, and this main effect was indeed qualified by an ity, mens rea) as a between-subjects variable and with each rating 2 interaction for punishment, F(3, 380) ϭ 3.40, p Ͻ .018, ␩p ϭ .026: as a dependent variable. No significant differences were found for ignorance of the vital fact was more mitigating for mens rea crimes harm, F(1, 394) ϭ 0.20, p Ͻ .66, d ϭ .04, 95% confidence interval than for strict liability crimes. (CI) [–.296, .469], anger, F(1, 394) ϭ 1.74, p Ͻ .19, d ϭ .13, 95% To confirm this interpretation statistically, we repeated our CI [–.126, .635], or disgust, F(1, 394) ϭ 1.35 p Ͻ .25, d ϭ .12, analysis with only the knowing and unknowing conditions (see 95% CI [–.169, .658], suggesting that the effects of legal category Figure 2), which revealed that for mens rea crimes, punishment we report below are unlikely to result from these potential con- ratings were significantly higher for the knowing condition than founds. for the unknowing condition, t(80) ϭ 6.24, p Ͻ .000, d ϭ 1.40, Moral censure and punishment. Do mental states play dif- 95% CI [1.48, 2.86] (corrected for violating Levene’s test), but that ferent roles in intuitive judgments concerning strict liability and for strict liability crimes, there was no significant difference be- mens rea crimes? And if so, does the relationship hold across tween the two conditions, t(98) ϭ 1.89, p Ͻ .061, d ϭ .38, 95% CI moral-purity and public welfare crimes? To address these ques- [–.033, 1.39]. The corresponding interaction for moral censure was tions we performed a series of 2 (Legal Category: strict liability, 2 not significant, F(3, 380) ϭ 2.33, p Ͻ .074, ␩p ϭ .018, but trended mens rea) ϫ 2 (Moral Classification: moral, public welfare) ϫ 4 in the same direction. These findings challenge Hypothesis 1b: (Mental State: knowing, unspecified, unknowing, deceived) Contrary to our initial expectations, and contrary to common ANOVAs for moral censure and punishment. assumptions in legal scholarship, our data support an intuitive These analyses revealed a significant main effect of mental state basis for the special treatment of strict liability crimes. Moreover, 2 for both moral censure, F(3, 380) ϭ 29.34, p Ͻ .000, ␩p ϭ .188, this pattern was robust across crime pairs: for each of our four 2 and punishment, F(3, 380) ϭ 20.52, p Ͻ .000, ␩p ϭ .139 (see pairs, and for both moral censure and punishment, the difference Figure 1). Independent samples t tests revealed that ratings for between the knowing and unknowing conditions was greater for each of these dependent variables were significantly higher in the the mens rea crime than for the strict liability crime, with a single knowing condition than in any of the other three conditions (ps Ͻ exception: for moral censure, the difference between ratings in the .003, ds Ͼ .43), consistent with most previous work on the effects knowing and unknowing conditions was the same for statutory of intent on moral judgments, and supporting Hypothesis 1. We rape as for seducing a minor. also found that ratings for the unspecified condition were signifi- Finally, for moral censure, there was a two-way interaction between mental state and moral classification, F(3, 380) ϭ 3.12, 2 p Ͻ .026, ␩p ϭ .024: for moral-purity crimes, moral censure did 7 not differ across the knowing and unspecified conditions, t(94) ϭ .548, p Ͻ .585, d ϭ .11, 95% CI [–.546, .963], nor across the

This document is copyrighted by the American Psychological Association or one of its allied6 publishers. unknowing and deceived conditions, t(98) ϭ .548, p Ͻ .548, d ϭ This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 5 .11, 95% CI [–.853, .455], but was significantly higher for both of the former conditions than for either of the latter conditions. In 4 contrast, for public welfare crimes, moral censure was signifi- 3 cantly higher for the knowing condition than all others, p Ͻ .001, d Ͼ .75, which did not differ from each other, ps ranged from .09 2 to .53, ds ranged from .13 to .35. Bases for punishment: Deterrence and retribution. To test 1 Moral Censure Punishment whether punishment was justified differently across legal categories, deterrence and retribution ratings were each subjected to a 2 Knowing Unspecified Unknowing Deceived (Legal Category: strict liability, mens rea) ϫ 4 (Mental State: Figure 1. Effect of mental state on moral censure and punishment. knowing, unspecified, unknowing, deceived) ANOVA. For deter- Higher values correspond to greater amounts of moral censure and pun- rence, this analysis revealed a main effect of legal category, F(1, 2 ishment. Error bars correspond to 1 SEM in each direction. 388) ϭ 9.50, p ϭ .002, ␩p ϭ .24, with higher ratings for mens rea WRONG OR MERELY PROHIBITED 711

Table 2 Means as a Function of Legal category

Dependent Experiment Measure Legal Category Knowing Unspecified Deceived Unknowing

Exp. 1 Moral censure Strict liability 4.78 (1.96) 4.50 (1.78) 2.82 (2.12) 3.44 (2.34) Mens rea 5.14 (1.93) 3.71 (2.06) 2.51 (1.67) 2.43 (1.80) Punishment Strict liability 3.96 (1.64) 3.60 (1.55) 2.71 (1.87) 3.28 (1.94) Mens rea 4.27 (2.07) 3.06 (1.82) 2.16 (1.18) 2.10 (1.30) Exp. S1 Moral censure Strict liability 4.04 (2.00) 4.04 (1.83) 3.50 (2.08) 3.71 (2.18) Mens rea 6.02 (1.30) 4.50 (1.81) 3.04 (2.20) 3.54 (2.12) Punishment Strict liability 4.29 (1.66) 4.21 (1.76) 3.44 (1.90) 4.15 (1.74) Mens rea 4.92 (1.57) 3.69 (1.53) 2.58 (1.67) 3.21 (1.88) Note. SD in parenthesis.

crimes (M ϭ 4.82, SD ϭ 1.88) than for strict liability crimes (M ϭ asignificantimpactonjudgments(supportingHypothesis1), 4.22, SD ϭ 1.95), contrary to Hypothesis 3 and what one might with ignorance having a mitigating effect (regardless of its expect on the basis of legal scholarship, which sometimes justifies source, challenging Hypothesis 1a), the effect of mental states the imposition of strict liability on utilitarian grounds related to was not uniform across legal category. Contrary to our predic- deterrence. This surprising result is unlikely to be a simple artifact tion (Hypothesis 1b), participants did differentiate between of our stimulus materials: recall that moral censure and punish- strict liability and mens rea crimes: ignorance was significantly ment were both lower for mens rea crimes than for strict liability less mitigating for strict liability crimes than for closely crimes, and we did not find main effects for anger, disgust, or harm matched “mens rea” crimes. across legal category. No significant results were found for retri- We also tested, but failed to find support for, three hypotheses 2 bution (ps Ͼ .215; ␩ps Ͻ .008). about why mental states might have different effects across legal Inferring knowledge and intent. Responses to the true/false categories. First, we considered the possibility (Hypothesis 2) that knowledge question were checked to confirm that participants mental states would be less important for intuitive judgments of correctly responded “true” for the knowing condition and “false” moral-purity crimes relative to public welfare crimes, consistent for the unknowing and deceived conditions. Accuracy was high, with both psychological data (Barrett et al., 2016; Russell & ranging from 93% for the deceived condition to 95% for the Giner-Sorolla, 2011) and the historical association between strict knowing condition and 99% for the unknowing condition. liability and purity (United States v. Morissette, 1952). While we For the unspecified condition, we were interested in whether did find that mental state interacted with moral classification— assumptions about knowledge would differ for mens rea and strict with the unspecified conditions patterning differently—we did not liability crimes. We therefore conducted a chi-squared test com- find the anticipated difference between the knowing and unknow- paring rates of inferred knowledge. This analysis revealed no ing conditions, nor an interaction with legal category. It thus difference, challenging prediction Hypothesis 4: in both cases, appears that violating purity is neither necessary nor sufficient to participants inferred knowledge 54% of the time. explain the attenuated role for mental states observed in strict liability cases. Discussion Second, we tested the idea (Hypothesis 3) that for strict liability In Experiment 1, we sought to establish whether moral and crimes, punishment may be warranted for reasons of deterrence, if punitive judgments concerning strict liability crimes reflect not retribution. However, we found no global differences in retri- their treatment by the legal system, according to which knowl- bution ratings, and lower deterrence ratings for strict liability edge is irrelevant. While we found that mental states did have crimes than for mens rea crimes. This is surprising given that This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Figure 2. Interaction between mental state and legal category for moral censure (not significant) and punishment (significant). Error bars correspond to 1 SEM in each direction. 712 GIFFIN AND LOMBROZO

punishment ratings were higher for strict liability crimes than for Experiment 2 mens rea crimes, even though crime types did not differ in per- Experiment 1 found support for the idea that laypeople’s intu- ceived harm. This is not to say that deterrence might not still itive judgments mirror the law in the following way: for mens rea furnish some justification for the imposition of strict liability, but crimes, judgments are sensitive to differences in mental states our data do suggest that for the crimes tested, laypeople do not (knowing vs. unknowing), while for strict liability crimes, judg- associate punishment with deterrence more strongly for strict ments are less sensitive (or even indifferent). The data also re- liability crimes than for mens rea crimes. vealed high levels of punishment assigned to strict liability offend- Finally, we also failed to find support for the idea (Hypothesis ers, despite the fact that strict liability offenses were judged no 4) that inferences about mental states might differ across legal worse than mens rea crimes (in terms of overall harm or elicited category, with mental states being more likely to be inferred (or anger and disgust), and despite the fact that three common justi- perhaps more difficult to ascertain) for strict liability: In our fications for punishment—retribution, deterrence, and incapacita- unspecified condition, rates for inferred knowledge did not differ tion (tested in S2)—were not endorsed more strongly for strict across strict liability and mens rea crimes. liability crimes than for mens rea crimes. In sum, the results of Experiment 1 suggest that laypeople’s Experiment 2 considers an explanation for these results with intuitive moral judgments mirror the law in reserving strict liability roots in the Supreme Court case of United States v. Morissette crimes for special treatment, but do not shed light on why this (1952). In Morissette, Justice Jackson noted that historically many might be. Before moving on to Experiments 2 and 3, in which we strict liability crimes were regulatory in nature, with the aim of find support for one possibility, it’s worth addressing some open improving public safety and welfare. He further suggested that questions from Experiment 1, which we investigated in supple- strict liability crimes were therefore primarily malum prohibitum mentary experiments. rather than malum in se, meaning that transgressions were consid- First, one concern is that the results from Experiment 1 could ered wrong because they were prohibited (i.e., violated a regula- be driven by the particular crimes that we happened to test. tion), not wrong or evil in themselves. For instance, while it isn’t Moreover, by selecting crimes that were so closely matched intrinsically wrong to drive a car at 50 mph, it is wrong to do so across legal categories (e.g., speeding paired with reckless in a 25-mph zone. To quote Justice Jackson’s Supreme Court driving), we could have inadvertently selected crimes that are opinion (United States v. Morissette, 1952), not representative of strict liability or mens rea crimes in general. To address this possibility, we ran an additional ex- While such offenses do not threaten the security of the state in the manner of treason, they may be regarded as offenses against its periment which we refer to as Experiment S1 (contact authors authority, for their occurrence impairs the efficiency of controls for a full report). In Experiment S1, we selected strict liability deemed essential to the social order as presently constituted. In this and mens rea crimes based on how frequently they are charged, respect, whatever the intent of the violator, the injury is the same... as represented by Federal Bureau of Investigation (FBI) crime Hence, legislation applicable to such offenses...does not specify statistics (FBI, 2012)andconsultationwithattorneys.Thestrict intent as a necessary element. (pp. 255–256) liability crimes selected were statutory rape, selling drugs The legal distinction between malum prohibitim and malum in within 1,500 ft of a school, speeding, and driving under the se mirrors a cognitive distinction explored in developmental re- influence. The mens rea crimes selected were sexual battery, search. Studies have found that children as young as six are able to burglary, theft, and (nonsexual) battery. Other than crime se- distinguish between actions that are “rule contingent,” or wrong lection, the methods and procedure mirrored those of Experi- because an authority figure or rule says they are wrong, and actions ment 1. Our key findings from Experiment 1 were replicated that are intrinsically wrong, and would therefore be wrong even if with these new crimes: ignorance significantly mitigated moral no rule or authority prohibited them (Nucci & Turiel, 2009; Turiel, censure and punishment ratings for mens rea crimes, but had no 2008). For example, if there is a rule against wearing pajamas to significant effect for strict liability crimes. Ratings for deter- school, then doing so might be wrong and merit detention, no rence were again higher for mens rea crimes. matter that the action would be perfectly acceptable on pajama A second concern is that the judgments concerning punishment day. Similarly, it could be that participants judged our strict This document is copyrighted by the American Psychological Association or one of its alliedin publishers. Experiment 1 may have been insufficiently fine-grained. In liability crimes wrong and deserving of punishment in large part This article is intended solely for the personal use ofsupplementary the individual user and is not to be disseminated broadly. Experiment S2 (contact authors for full report), because they violated rules, even though they did not regard the participants separately specified the appropriate levels for fines actions themselves as more harmful, angering, or disgusting than and for jail time. We also included a third possible basis for the mens rea crimes, and even though retribution, deterrence, and punishment: incapacitation. Experiment S2 supported the interpre- incapacitation were rated more appropriate reasons for punishing tation of Experiment 1: we found the same pattern of responses commissions of mens rea crimes. concerning the role of mental states across legal categories, and we To test these ideas, participants evaluated vignettes written in found that incapacitation was rated a significantly better basis for the style of our unspecified condition from Experiment 1. Then, punishing mens rea crimes than strict liability crimes. subsequent to reading and evaluating the original vignettes, par- In conjunction with the findings from Experiment 1, these ticipants were told that the relevant statute had been changed. For supplementary experiments provide support for the idea that peo- example, participants in the statutory rape vignette were told, after ple mirror the law in their treatment of mens rea versus strict initial judgments, that the age of consent had been lowered to 15, liability crimes, but the findings do not (yet) reveal the basis for meaning that the action described in the vignette was no longer a this special treatment. crime. Participants then provided a second set of ratings for moral WRONG OR MERELY PROHIBITED 713

censure and punishment, allowing us to assess how much the Jail time: How much jail time should [Actor] receive? change in statute affected these judgments, and therefore how “rule contingent” (or malum prohibitum) they regarded the original Participants indicated their answers on a continuous slider that wrong. ranged from 0 to twice the maximum penalty that could actually be Finally, we also added a question about whether the offender assigned for the most harshly punished crime, assuming a first had a responsibility to acquire the knowledge relevant to the crime offense. The maximum value for jail time was 36 years. (e.g., the speed at which he was driving). If people believe that Fines: How much should [Actor] be fined? strict liability crimes have higher demands in this regard, then ignorance may be treated as a form of negligence, deserving of Participants indicated their answers on a continuous slider that punishment despite an absence of bad intent. ranged from 0 to twice the maximum penalty that could actually be In sum, Experiment 2 was designed to test two hypotheses. The assigned for the most harshly punished crime, assuming a first central hypothesis (Hypothesis 5) was that punishment and moral offense. The maximum value for fines was $100,000. We did not censure ratings would decrease after participants learned of the ask the deterrence, retribution, or incapacitation questions in this statute change (such that the action was no longer prohibited), and experiment. that this decrease would be greater for strict liability crimes than We added two new dependent variables about the actor’s knowl- for mens rea crimes. A secondary hypothesis (Hypothesis 6) was edge and responsibility to acquire that knowledge. Examples from that participants would believe offenders had a greater responsi- the speeding story are below: bility to avoid ignorance for strict liability crimes than for mens rea crimes. Inference: Did Alan know that he was driving over the speed limit? Method This was rated from 1 (no, definitely not) to 7 (yes, definitely).

Participants. Four hundred thirteen adults (257 male, 154 Responsibility: Did Alan have a responsibility to check female, two other/prefer not to specify, Mage ϭ 31, SD ϭ 10) whether he was going over the speed limit? participated in the study through Amazon Mechanical Turk. An additional 77 participants were tested, but were excluded for Rated from 1 (no, definitely not) to 7 (yes, definitely). failing catch questions. Only U.S. participants who indicated After participants answered these questions, we presented them they were older than 18 were allowed to participate, to approx- with a statute change that stated that the law had been changed imate a real jury pool. Participants were given $0.50 for their such that the actor’s action was no longer illegal. In each case, we participation. were clear that the statutory change was due to a technicality, not Materials and procedures. In an effort to view strict liability due to any change in the harm caused by the action. The partici- and mens rea crimes broadly, we included all the crimes used in pants then provided moral censure and punishment ratings in light Experiments 1 and S1 and added three new strict liability and three of the new statutory information. An example from the migratory new mens rea crimes. The new strict liability crimes were illegal bird story is below: dumping, possession of an unregistered firearm, and hunting a migratory bird. The new mens rea crimes were disturbing the After Alan’s arrest, the officer was informed that the statute had been changed. Due to a technicality (not to a change in the bird species’ peace, public drunkenness, and minor in possession of alcohol (see status), the bird Alan shot was no longer on the list of protected birds. Table 3). As in Experiment S1, these new crimes were chosen as Alan’s actions were therefore no longer technically a crime. Please the next most frequently charged according to FBI statistics and answer the following questions based on the change in statute. consultation with attorneys. The resulting design had 20 conditions, one for each crime. One concern with including this statute change as a within- After reading a vignette describing a crime, participants were subjects factor is that participants’ initial ratings (before the statute asked the moral censure, disgust, harm, and anger questions used change) could anchor or otherwise influence their second ratings in Experiment 1. Participants were also asked two punishment (after the statute change). We were not overly worried by this questions—about fines and jail time: possibility because our key hypothesis (Hypothesis 6) concerned a This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Table 3 Crimes Used in Experiments 2 and 3 in Addition to Those Used in Experiment 1

Strict liability Mens rea

Illegal dumping (distance from waterway) Disturbing the peace (unlawful to fight in public) Unregistered firearm (registration status of firearm) Public drunkenness (intoxication level prevented taking care) Hunting migratory bird (bird’s protected status) Minor in possession (beverage contained alcohol) Selling drugs within 1,500 ft (distance from nearest school) Sexual battery (consent for the touching) Driving under the influence (level of intoxication) Burglary (owner’s intent to keep the vase) Theft (whether saw had been discarded) Battery (whether friend was genuinely in danger) Note. Each crime is followed in parentheses by the relevant knowledge that was varied across the mental state manipulation in Experiment 3. 714 GIFFIN AND LOMBROZO

differential drop in ratings across strict liability versus mens rea The relationship between legal category and moral censure re- crimes, not the values of the ratings themselves. But if anything, mained significant, r ϭϪ.165, p Ͻ .001, as did that for legal this methodological choice works against our hypotheses. category and fines, r ϭϪ.126, p ϭ .012. Analyses across crimes. Not surprisingly, there was consid- Results erable variation across crimes for all of our measures (see Appen- dix A for additional data by crime). Figure 3 plots the moral Moral censure and punishment: Ratings under statutory censure difference scores for each crime. The figure makes it clear change. To test Hypothesis 5, that the wrong associated with that while the means for mens rea and strict liability crimes strict liability crimes is more contingent on the presence of a rule differed, their distributions were overlapping. In the general dis- than that of mens rea crimes, we created difference scores for cussion we return to some of the outlying cases. moral censure, fines, and jail time by subtracting each participant’s rating after the statute change from that participant’s correspond- Discussion ing initial rating. An independent samples t test on moral censure difference scores revealed a significant difference in the predicted Experiment 2 had several aims. Most importantly, we found direction, t(411) ϭ 3.16, p Ͻ .002, d ϭ .31, 95% CI [.155, .665]: support for one account of what differentiates strict liability crimes ratings dropped an average of .914 points (SD ϭ 1.41) on a 7-point from mens rea crimes. For the crimes we tested, the mere act of scale for strict liability crimes, but only .504 points (SD ϭ 1.23) violating a prohibition played a larger role in judgments of moral for mens rea crimes. censure and punishment for strict liability crimes than for mens rea To analyze punishment ratings, we transformed fines and jail crimes. Specifically, a change in statute led to a significantly time differences into z scores and computed a repeated measures greater decrease in moral censure and fines for strict liability ANOVA with punishment type as a within-subjects factor and crimes than for mens rea crimes, supporting Hypothesis 5. This is legal category as a between subjects factor. While the main effect consistent with the idea that strict liability crimes are wrong to a of legal category was not significant, F(1, 398) ϭ 2.26, p ϭ .103, greater extent because they are prohibited; they are more malum 2 prohibitum than mens rea crimes, which are, to a greater extent, ␩p ϭ .007, there was a nearly significant interaction between punishment type and legal category, F(1, 398) ϭ 3.70, p ϭ .055, malum in se. 2 Experiment 2 also found a difference in the knowledge imputed ␩p ϭ .009. For fines only, the predicted effect was significant, t(303) ϭ 2.36, p Ͻ .019, d ϭ .28, 95% CI [449.9, 496.9] (corrected to offenders of different crimes: the modal response for strict for violating Levene’s test): for strict liability crimes, there was a liability crimes was to infer an absence of knowledge; the modal fine drop of $4,844 (SD ϭ $13,595), versus only $2,134 (SD ϭ response for mens rea crimes was the midpoint of our scale. This $8,637), for mens rea crimes. Jail time also dropped a larger difference does not appear to have driven the greater “rule con- amount for strict liability crimes (M ϭ 0.364, SD ϭ 1.09) than for tingency” of strict liability crimes. However, it may nonetheless be mens rea crimes (M ϭ 0.339, SD ϭ 1.35), but not significantly, t(411) ϭ .198, p Ͻ .843, d ϭ .03, 95% CI [Ϫ.217, .266]. Drug Distribuon Inferred knowledge and responsibility. Responses to the Mens Rea question about the actor’s responsibility for obtaining reliable First Cousins information differed significantly between strict liability (M ϭ The Strict Liability 5.23, SD ϭ 2.04) and mens rea crimes (M ϭ 5.61, SD ϭ 1.83) Burglary crimes, t(378) ϭ 1.98, p Ͻ .049, d ϭ .20, 95% CI [Ϫ.759, Ϫ.002] Baery (corrected for violating Levene’s test), with higher responsibility Incest in the mens rea case. If anything, this goes counter to our expec- Unregistered Firearm tation (Hypothesis 6) that people might be held more responsible Disturbing the Peace in strict liability cases, rendering ignorance akin to a form of Public Drunkenness negligence deserving punishment. Sexual Baery We also found that participants were significantly more likely to Drugs to Minors believe that the principal actor possessed the relevant knowledge 1,500 Feet This document is copyrighted by the American Psychological Association or one of its allied publishers. for mens rea crimes (M ϭ 3.74, SD ϭ 2.04) than for strict liability Minor in Possession This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. crimes (M ϭ 2.61, SD ϭ 1.62), t(410) ϭ 6.27, p Ͻ .000, d ϭ .62, Seducing a Minor 95% CI [Ϫ1.48, .774]. This differs from what we found in Exper- Dumping iment 1 in relation to Hypothesis 4; in that experiment, we found DUI no differences in inferred knowledge across mens rea and strict Reckless Driving liability crimes. It is possible that the difference revealed here is Protected Bird the result of a more sensitive, continuous measure, or to the Statutory Rape inclusion of additional crimes. Whatever the cause, it raises the Speeding question of whether the greater drop in ratings for strict liability crimes than for mens rea crimes after the statutory change was a 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 consequence of differences in inferred knowledge, or an indepen- Moral Censure Diﬀerence Score dent effect. To test this, we computed partial correlations between legal category and each of our three difference scores (moral Figure 3. Moral censure difference scores (rating before statute change censure, fines, and jail time), controlling for inferred knowledge. minus rating after statute change) by crime in Experiment 2. WRONG OR MERELY PROHIBITED 715

related. To the extent that strict liability crimes hinge on the tested, but were excluded for failing catch questions. Only U.S. precise specification of a rule (e.g., whether a bird or substance is participants who indicated they were older than 18 were allowed to on an official list), ignorance may be plausible, yet still fail to participate, to approximate a real jury pool. Participants were mitigate: because the crimes are malum prohibitum, the guilty given $0.65 for their participation. mind is less relevant. Materials and procedures. Participants were randomly assigned to one of 20 conditions, representing each of the 20 crimes Experiment 3 from Experiment 2. Each participant initially read the unknowing vignette of the assigned crime. Most of the unknowing versions Experiment 1 found that ignorance was less mitigating for strict had been used in previous experiments, but for the crimes added in liability crimes than for mens rea crimes, but failed to identify a Experiment 2, new unknowing versions were created (see online potential basis for this differential effect. In Experiment 2, we supplement for the six new vignettes). After reading their unknow- found support for one relevant difference between strict liability ing vignette, participants were asked the moral censure, fine, and and mens rea crimes: on average, the former were treated as more jail time questions used in Experiment 2. “rule contingent” in the sense that censure and punishment Next, participants were presented with a knowledge change. dropped to a greater extent after a rule change. This is consistent Participants were asked to imagine that the actor had in fact known with the suggestion that strict liability crimes are somewhat akin to the vital fact, and they then provided the moral censure, fine, and conventional violations, or malum prohibitum. A remaining ques- jail ratings again. This manipulation allowed us to obtain a within- tion, however, is why strict liability is treated in this way. The subjects measure of the effect of mental states. An example from historical examples cited by Justice Jackson in United States v. the migratory bird vignette is presented below: Morissette (1952) were largely public safety regulations, but many contemporary strict liability crimes address other domains of hu- Suppose that Alan had actually realized, when he shot the bird, that his actions broke the law—that is, that he was shooting a protected bird. And man behavior. What is it, then, about strict liability crimes that suppose that he decided to shoot the bird anyway. In this case, where renders their evaluation more contingent on the presence of a rule, Alan knowingly violated the law, how would you respond to the follow- and less contingent on mens rea? ing questions? (Your responses may be the same as those you just While not all contemporary strict liability crimes involve public provided, or they may differ.) safety or welfare regulations, many seem to involve a somewhat arbitrary element. Returning to an earlier example, it is not in Next, all participants, regardless of which crime they initially itself wrong or illegal to drive a car 50 mph, but it is illegal to do evaluated, rated the arbitrariness of all 20 crimes. The crimes were so in a 25-mph zone. But why is the zone’s speed limit designated preceded by the following instructions: as 25 mph, and not 24 or 26? Similarly, it is not in itself wrong or Below you will see a list of laws. We’d like you to give your intuitions illegal to have sex, but in our country it is illegal for an adult to do about how arbitrary each law seems to you. That is, do you believe that so with a person who is below a certain age. But why is that age there’s a good reason for the law to draw the line where it does in terms 18 in some states, and 16 in others? Of course, there are very good of which actions are legal versus illegal? Or does it seem like the law is reasons why driving too quickly or having sex with a child could somewhat arbitrary in the sense that it could reasonably have drawn the be considered wrong in themselves—the rules are not merely a line differently? matter of convention. But the specific line that is drawn between legal and illegal actions—the specific age, speed, distance, and so Please do not consult any outside resources, like other people or forth, specified in a prohibition—is somewhat arbitrary and may websites. We are interested in your own intuitions, whether or not they correspond to the current legal system. Even if you think all laws therefore render strict liability crimes a hybrid of malum in se and are somewhat arbitrary or not at all arbitrary, please take note of malum prohibitum, whereas most mens rea crimes are more which seem more or less arbitrary and respond accordingly. cleanly malum in se. To test this idea, Experiment 3 investigated whether participants The 20 crimes were then presented in a randomized order. Each consider strict liability crimes to be more arbitrary than mens rea crime was presented as true to its statutory language as possible, crimes. Our first prediction (Hypothesis 7) was that, on average, with minor adjustments in wording to facilitate comprehension. the rules corresponding to strict liability crimes would be judged Providing the full statutory language ensured that participants were This document is copyrighted by the American Psychological Association or one of its allied publishers. more arbitrary than those corresponding to mens rea crimes. We not operating with different, common sense definitions of each This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. also obtained a measure reflecting the role of mental states for each crime. Below is the text for migratory bird: crime by presenting unknowing and knowing conditions to each participant. We predicted that the effect of mental states would be Migratory birds: It is illegal for anyone to pursue, hunt, take, smaller for strict liability crimes than for mens rea crimes. We also capture, or kill any migratory bird, or any part, nest, or egg of any predicted an association between arbitrariness ratings and the such bird. magnitude of the effect of mental state (Hypothesis 8): The more Participants indicated their agreement on a scale of 1 (not at all arbitrary the rule, the smaller the influence of mens rea. arbitrary)to7(completely arbitrary). Finally, participants answered the demographic and catch questions used in previous experiments. Method Results Participants. Four hundred four adults (183 male, 221 fe-

male, Mage ϭ 35, SD ϭ 12) participated in the study through Arbitrariness ratings. A primary aim of Experiment 3 was to Amazon Mechanical Turk. An additional 80 participants were test whether strict liability crimes are considered more arbitrary 716 GIFFIN AND LOMBROZO

than mens rea crimes (Hypothesis 7). To do so, we computed the crimes are not wrong in themselves, but merely that many also average arbitrariness rating for strict liability crimes and for mens have a conventional element, and that the role for such arbitrary or rea crimes for each participant. We then ran a paired-samples t test conventional elements is greater, on average, for strict liability on mean arbitrariness ratings as a function of legal category. crimes than for mens rea crimes. Consistent with our prediction, participants rated strict liability We also found a significant, though small, correlation between crimes significantly more arbitrary (M ϭ 2.99, SD ϭ 1.61) than participants’ arbitrariness ratings and the extent to which their mens rea crimes (M ϭ 2.93, SD ϭ 159), t(403) ϭ 1.87, p Ͻ .031, judgments varied across knowing and unknowing violations. Spe- d ϭ .04, 95% CI [–.122, .003] (one-tailed test). (Mean arbitrariness cifically, the perpetrator’s mental states (mens rea) had a greater ratings for each crime can be found in Appendix B.) impact on ratings the less arbitrary the rule. We return to this Moral censure and punishment: Ratings under mental state relationship in the general discussion. change. This experiment was the first in which we manipulated mental states within subjects. Thus, we were able to create differ- General Discussion ence scores for each participant for moral censure, fines, and jail time by subtracting each participant’s initial, unknowing ratings In the work reported here, we set out to test whether the legal from the rating given after participants were asked to assume that distinction between strict liability and mens rea crimes reflects the perpetrator acted knowingly. An independent samples t test (and potentially reveals) aspects of intuitive moral judgment. Here, comparing moral censure difference scores as a function of legal we summarize and discuss our key findings. category failed to find a significant effect, t(402) ϭ .126, p Ͻ .899, Experiment 1 revealed that judgments concerning mens rea d ϭ .01, 95% CI [–.356, 406]. This departure from our previous crimes were in fact more sensitive to the actor’s knowledge and findings could be due to the within-subjects design or to the intent than were those concerning strict liability crimes. When an addition of six crimes that were not included in Experiment 1 or in actor transgressed knowingly, punishment and censure were the supplementary experiments, and which were much less famil- greater for mens rea crimes than for strict liability crimes, but iar than those used in our initial studies. We ran the same inde- when the actor transgressed without knowledge and intent pendent samples t test on difference scores for the original 14 (whether merely ignorant or deceived), this pattern was reversed. crimes from Experiments 1 and S1 and found a significantly As a result, judgments for mens rea crimes were strongly influ- greater effect of mental state for mens rea than for strict liability enced by the actor’s knowledge and intent, whereas those for strict crimes on moral censure, t(273) ϭ 4.08, p Ͻ .000, d ϭ .49, 95% liability crimes were about the same. Put differently, ignorance CI [.475, 1.36] (corrected for violating Levene’s test), replicating was treated as a mitigating factor for mens rea crimes, but not for the original effect. strict liability crimes. For punishment, however, we did find the predicted effect when In Experiments 2 and 3, we found support for a possible expla- considering all 20 crimes. A repeated measures ANOVA with nation: that strict liability crimes have more features that are legal category as a between subjects factor, and z scores of differ- malum prohibitum, or wrong because prohibited, while mens rea ence scores for fines and jail time as a within subjects fact or, crimes are largely malum in se, or wrong in themselves. Experi- revealed a significant effect of legal category, F(1, 398) ϭ 3.88, ment 2 found that judgments for strict liability crime were more 2 p ϭ .050, ␩p ϭ .008. dependent on the presence of a rule: when a statute was revoked, Correlation between mental state effects and arbitrariness ratings for moral censure and punishment decreased more for strict of assigned story. Finally, we tested whether the arbitrariness of liability crime than for mens rea crimes. These findings have a crime would relate to the magnitude of a mental state effect for analogues in social and cognitive development, where research that crime. If part of what differentiates strict liability crimes from differentiates between moral and “conventional” wrongdoing (Nu- mens rea crimes is their more conventional or arbitrary nature, and cci & Turiel, 2009; Turiel, 2008). Experiment 3 focused on another this difference is partially responsible for the relative unimpor- aspect of conventionality: the potentially arbitrary nature of the tance of mens rea, then we would expect arbitrariness ratings to be rule involved. We found that rules prohibiting strict liability crimes negatively correlated with the magnitude of a mental state effect. were on average judged more arbitrary than rules for mens rea We ran a bivariate correlation between each participant’s knowing- crimes, and that the more arbitrary the rule, the more attenuated the unknowing difference score and her arbitrariness rating for the role of mental states. This document is copyrighted by the American Psychological Association or one of its allied publishers. specific crime she read. Consistent with our prediction (Hypothesis The relationship between strict liability and arbitrariness can This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 8), we found that moral censure scores were significantly nega- help explain why strict liability crimes involve an attenuated role tively related to arbitrariness, r(404) ϭϪ.13, p Ͻ .010. We found for mental states. First, the more arbitrary nature of strict liability no significant correlations with either punishment measure. crimes means that the intentions associated with a knowing violation are not necessarily intrinsically bad. For instance, intention- Discussion ally driving at 55 mph is not itself wrong; it only becomes so in the context of a law that specifies an area as a 50-mph zone. When a Experiment 3 confirmed our prediction (Hypothesis 7) that strict law is not arbitrary, by contrast, the intention is reprehensible even liability crimes are considered more arbitrary than mens rea absent a law: intentionally hitting another person is unacceptable crimes. This prediction was motivated in part by arguments de- whether or not a law is in place. Relatedly, crimes that involve veloped in Morissette and the findings from Experiment 2: If an arbitrary thresholds are likely to impose a looser coupling between action is wrong merely because it is prohibited, then the prohibi- the legal status of an action and the “wrongness” of the act. tion must not be grounded in intrinsic harm or other moral con- Dumping a chemical a few feet outside a recognized zone, for siderations. To be clear, our contention is not that strict liability example, could be regarded as only slightly less harmful than WRONG OR MERELY PROHIBITED 717

dumping the chemical just inside the zone. Yet crossing the thresh- of the human will and a consequent ability and duty of the normal old to illegal status will make the act wrong not just by an individual to choose between good and evil. (p. 250) incremental increase in harm, but by virtue of having violated a prohibition. For mens rea crimes, in contrast, harm may track legal In contrast, we found that our participants’ judgments were status more closely: taking something that doesn’t belong to you is surprisingly consistent with the law, finding mental states less both wrong and illegal; not doing so is neither. For these reasons, relevant for strict liability crimes than for other crimes. We are not the evaluation of strict liability crimes—relative to mens rea arguing that mental states play no role for strict liability crimes, but crimes—could be less sensitive to mental states, and also more our results do suggest that the (putative) counterintuitiveness of dependent on the presence of a rule. this aspect of strict liability has been overstated, and might not be In light of these considerations, we can revisit our specific the most compelling basis for arguments by its opponents. Indeed, crimes. Reckless driving, seducing a minor, and minor in posses- recent work in the realm of torts has found that laypeople may sion—all mens rea crimes—also involve a potentially arbitrary endorse strict liability more strongly than scholars (Sanders, Ku- threshold or contingency. And in fact, in Experiment 2, reckless gler, Solan, & Darley, 2014). driving had the greatest drop in moral censure after the statute Our studies also bear on more general questions about the role change of any mens rea crime, while seducing a minor had the of explanations in legal judgment. We found evidence that offering second greatest drop in moral censure and the highest for fines and an explanation for an actor’s mental states can be more mitigating jail time of any mens rea crime (see Figure 3 and Appendix A). than the mere absence of knowledge. Specifically, Experiments 1 Minor in possession had the third greatest reduction in moral and S1 found that participants in the unspecified condition gave censure and a complete erasure of fines. In Experiment 3, minor in significantly higher ratings on every dependent measure than par- possession was also judged the second most arbitrary crime. On ticipants in either of the conditions that offered explanations for the other hand, the strict liability crimes of incest and unregistered false beliefs: unknowing or deceived. Experiment S1 also found firearm showed the smallest drops in moral censure in response to that deception was more mitigating than bad information. Our a statute change, and both arguably lack an arbitrary threshold studies thus go beyond prior work about the importance of knowl- beyond which they become legal. edge and intent by showing that while mental states are undeniably Considering these “exceptions to the rule” helps make a valu- important, explanations for those mental states are as well. able point. We do not mean to suggest that strict liability crimes form a special psychological kind, or that they have clear-cut necessary and sufficient conditions. Rather, a variety of different Limitations and Future Directions factors that come in degrees may be more likely to be found in One major concern with the present research is its lack of strict liability crimes, and these features can potentially help ex- ecological validity. Certainly, our participants were making judg- plain their legal origins as well as the way they’re evaluated by ments in a much less formal and demanding environment than both experts and laypeople. The surprise, perhaps, is that the strict actual jurors judging a real case. Moreover our participants did not liability designation has any counterpart in intuitive moral judg- engage in group deliberation, and had less information than would ments; that the match is imperfect is to be expected, especially be available at trial. However, these limitations do not undercut the given the recognized heterogeneity of strict liability crimes within central aim of the present research, which was to investigate the law (Levenson, 1992; Sayre, 1933). whether laypeople’s intuitive moral judgments discriminate be- One lingering concern is that our findings could reflect Amer- tween strict liability and mens rea crimes in the absence of specific ican citizens’ knowledge of their own legal system, and not their instructions to do so, and if so, why. untutored moral intuitions. Arguing against this possibility, an additional study confirmed that participants sampled from the Our results support the idea that the relative arbitrariness of a same population as that in our studies were no more likely to think crime, and therefore its status as malum prohibitum, is intuitively that knowledge was required for the conviction of our mens rea recognized and influences judgments. However, open questions crimes than of our strict liability crimes (see online supplement for remain concerning how participants were interpreting and evalu- details). ating arbitrariness. In addition, the relationship between arbitrariness and the role of mental states was significant, but small. This This document is copyrighted by the American Psychological Association or one of its alliedImplications publishers. for the Law suggests that a variety of additional factors influence the relative This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. importance of mental states in legal judgment, and identifying Finding that the legal category of strict liability mirrors a cog- these factors is an important direction for future research. nitive distinction is exciting, but also potentially surprising: some Another concern is the number of participants who failed our of strict liability’s most vocal criticism is rooted in the belief that attention check questions. Using an Internet sample necessitates defendants’ mental states should and would have an effect on their stringent attention checks, as participants cannot be monitored as legal outcomes. That is, there is often an implicit assumption that they would be in a lab. Despite our multiple checks, however, our jurors care about a defendant’s knowledge or intent, and that exclusion rates were comparable to those for other studies run on presenting evidence for the absence of either would mitigate a Mechanical Turk (e.g., Downs, Holbrook, Sheng, & Cranor, defendant’s penalty in strict liability cases. For instance, Justice 2010). One advantage of using a Mechanical Turk sample is that Jackson in United States v. Morissette (1952) said, it is more representative of a jury than a sample of university The contention that an injury can amount to a crime only when undergraduates. That said, extending this line of research to jury- inflicted by intention is no provincial or transient notion. It is as eligible participants who engage in deliberation over more detailed universal and persistent in mature systems of law as belief in freedom and realistic cases has clear value. 718 GIFFIN AND LOMBROZO

Conclusion Malle, B. F., & Knobe, J. (1997). The folk concept of intentionality. Journal of Experimental Social Psychology, 33, 101–121. http://dx.doi In sum, our studies are the first (to our knowledge) to relate the .org/10.1006/jesp.1996.1314 strict liability designation to laypeople’s intuitive moral judg- Malle, B. F., & Nelson, S. E. (2003). Judging mens rea: The tension ments. We find that strict liability and mens rea crimes differ in between folk concepts and legal concepts of intentionality. Behavioral two potentially related ways: mens rea crimes are more sensitive to Sciences & the Law, 21, 563–580. http://dx.doi.org/10.1002/bsl.554 the presence or absence of relevant knowledge, but strict liability Manchester, C. (1977). The origins of strict criminal liability. The Anglo- crimes are more contingent on the presence of a rule. Our studies American Law Review, 6, 277–283. Mens Rea: The Need for a Meaningful Intent Requirement in Federal suggest that people find strict liability crimes to be more arbitrary Criminal Law: Hearing before the Over-Criminalization Task Force of and malum prohibitum, while mens rea crimes appear to be, to a the Committee on the Judiciary, House of Representations, 113th Cong. greater extent, malum in se. This is good news for those, like (2013). Justice Holmes, who believe that the law should reflect community Mikhail, J. (2009). Is the prohibition of homicide universal? Evidence from standards, but it may still be troubling for those who believe the comparative criminal law. Brooklyn Law Review, 75, 497–516. law should maintain the highest standard of justice, even when the Murray, D., & Lombrozo, T. (2016). Effect of manipulation on attributions community does not demand it. of causation, free will, and moral responsibility. Cognitive Science: A Multidisciplinary Journal. Advance online publication. http://dx.doi.org/ 10.1111/cogs.12338 References Nucci, L., & Turiel, E. (2009). Capturing the complexity of moral development and education. Mind, Brain, & Education, 3, 151–159. Alicke, M. (2008). Blaming badly. Journal of Cognition and Culture, 8, Oppenheimer, D. M., Meyvis, T., & Davydenko, N. (2009). Instructional 179–186. http://dx.doi.org/10.1163/156770908X289279 manipulation checks: Detecting satisficing to increase statistical power. Barrett, H. C., Bolyanatz, A., Crittenden, A. N., Fessler, D. M., Fitzpatrick, Journal of Experimental Social Psychology, 45, 867–872. http://dx.doi S., Gurven, M.,...Scelza, B. A. (2016). Small-scale societies exhibit .org/10.1016/j.jesp.2009.03.009 fundamental variation in the role of intentions in moral judgment. Phillips, J., & Shaw, A. (2015). Manipulating morality: Third-party inten- Proceedings of the National Academy of Sciences of the United States of tions alter moral judgments by changing causal reasoning. Cognitive America, 113, 4688–4693. http://dx.doi.org/10.1073/pnas.1522070113 Science: A Multidisciplinary Journal, 39, 1320–1347. http://dx.doi.org/ Blackstone, W. (1769). Commentaries on the laws of England: Public 10.1111/cogs.12194 wrongs. Oxford, United Kingdom: Clarendon Press. Rai, T. S., & Fiske, A. P. (2011). Moral psychology is relationship Craswell, R., & Calfee, J. (1986). Deterrence and uncertain legal standards. regulation: Moral motives for unity, hierarchy, equality, and propor- Journal of Law Economics and Organization, 2, 279–314. tionality. Psychological Review, 118, 57–75. http://dx.doi.org/10 Downs, J. S., Holbrook, M. B., Sheng, S., & Cranor, L. F. (2010, April). .1037/a0021867 Are your participants gaming the system? Screening Mechanical Turk Russell, P. S., & Giner-Sorolla, R. (2011). Moral anger, but not moral workers. Proceedings of the SIGCHI Conference on Human Factors in disgust, responds to intentionality. Emotion, 11, 233–240. http://dx.doi Computing Systems (pp. 2399–2402). New York, NY: ACM. .org/10.1037/a0022598 Federal Bureau of Investigation. (2012). Crime in the United States. Sanders, J., Kugler, M., Solan, L., & Darley, J. (2014). Must torts be Uniform Crime Reports. Retrieved from https://ucr.fbi.gov/crime-in-the- wrongs? An empirical perspective. Wake Forest Law Review, 49, 1–45. u.s/2012/crime-in-the-u.s.-2012 Sayre, F. (1933). Public welfare offenses. Columbia Law Review, 33, Graham, J., Nosek, B. A., Haidt, J., Iyer, R., Koleva, S., & Ditto, P. H. 55–89. http://dx.doi.org/10.2307/1115413 (2011). Mapping the moral domain. Journal of Personality and Social Singer, R. (1989). The resurgence of mens rea: The rise and fall of strict Psychology, 101, 366–385. http://dx.doi.org/10.1037/a0021847 criminal liability. Boston College Law Review Boston College Law Hart, H. L. A., & Honore, A. M. (1962). Causation in the law (1st ed.). School, 30, 337–408. Oxford, United Kingdom: Oxford University Press. Stahlkopf, C., Males, M., & Macallair, D. (2010). Testing incapacita- Holmes, O. W. (1881). The common law. Boston, MA: Little, Brown, and tion theory: youth crime and incarceration in California. Company. Crime and Delinquency, 56, 253–268. http://dx.doi.org/10.1177/ Inbar, Y., Pizarro, D. A., & Cushman, F. (2012). Benefiting from misfor- 0011128707307227 tune: When harmless actions are judged to be morally blameworthy. Stanton-Ife, J. (2007). Strict liability: Stigma and regret. Oxford Jour- Personality and Social Psychology Bulletin, 38, 52–62. http://dx.doi nal of Legal Studies, 27, 151–173. http://dx.doi.org/10.1093/ojls/ .org/10.1177/0146167211430232 gql033

This document is copyrighted by the American Psychological Association or one of its alliedKeating, publishers. G. (2006). Strict liability and the mitigation of moral luck. Turiel, E. (2008). Thought about actions in social domains: Morality, social

This article is intended solely for the personal use of the individual userJournal and is not to be disseminated broadly. of Ethics & Social Philosophy, 2, 1–33. conventions, and social interactions. Cognitive Development, 23, 136– Knobe, J. (2005). Theory of mind and moral cognition: Exploring the 154. http://dx.doi.org/10.1016/j.cogdev.2007.04.001 connections. Trends in Cognitive Sciences, 9, 357–359. http://dx.doi.org/ United States v. Morissette, 342 U.S. 246 (1952). 10.1016/j.tics.2005.06.011 Young, L., & Saxe, R. (2011). When ignorance is no excuse: Different Levenson, L. (1992). Good faith defenses: Reshaping strict liability crimes. roles for intent across moral domains. Cognition, 120, 202–214. http:// Cornell Law Review, 78, 401–469. dx.doi.org/10.1016/j.cognition.2011.04.005 WRONG OR MERELY PROHIBITED 719

Appendix A Difference Scores and Mental State Correlations by Crime from Experiment 2

Moral censure Fines Jail Mental state and moral Mental state and Mental state and Crime difference score difference score difference score censure correlation fines correlation jail correlation

Speeding 1.65 273.06 .000 .403 .051 –.019 Statutory rape 1.60 4,645.07 .945 .031 .298 N/A 092. 415. ءMigratory bird 1.45 4,580.59 .255 .481 Reckless driving 1.30 –1,600.95 .130 .067 .104 .091 Driving under the influence .85 3,330.82 .360 .187 .025 .13 ء532. 192. ءءDumping .85 10,094.39 .285 .666 Seducing a minor .65 13,246.42 1.395 .083 .031 Ϫ.195 Minor in possession .60 116.40 .000 .112 Ϫ.384 N/A 1,500 ft .55 6,143.78 .596 Ϫ.086 Ϫ.047 Ϫ.084 Drugs to minors .55 6,264.11 .435 .364 Ϫ.148 .296 Sexual battery .55 3,560.32 .965 .402 .356 .233 ءPublic drunkenness .55 54.49 .005 .301 .100 .515 ء449. 141. ءDisturbing the peace .50 683.41 .070 .484 Unregistered firearm .45 6,421.05 .410 .137 .420 .304 Incest .30 307.06 .090 .111 .168 .128 Battery .30 1,526.40 .185 .054 Ϫ.164 .099 Ϫ.014 .132 ءBurglary .25 859.47 .300 .549 39. 207. ءTheft .20 614.03 .120 .560 First cousins .20 2,326.00 .265 .211 Ϫ.064 Ϫ.099 ء524. ء493. ءءDrug distribution .15 716.17 .220 .698 Note. N/A ϭ not applicable. .p Ͻ .01 ءء .p Ͻ .05 ء

(Appendices continue) This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 720 GIFFIN AND LOMBROZO

Appendix B Arbitrariness by Crime from Experiment 3

Crime Arbitrariness rating

Theft 2.37 (2.06) Burglary 2.38 (2.15) Sexual battery 2.54 (2.20) Battery 2.61 (2.12) Dumping 2.63 (2.07) Reckless driving 2.74 (2.06) Driving under the influence 2.82 (2.07) Seducing a minor 2.84 (2.15) Drug distribution 2.87 (2.07) Unregistered firearm 2.87 (2.14) Drugs to minors 3.03 (2.15) Speeding 3.03 (1.90) Migratory birds 3.05 (1.83) Incest 3.09 (2.21) Statutory rape 3.17 (2.14) 1,500 ft 3.19 (2.23) Disturbing the peace 3.20 (1.84) Public drunkenness 3.43 (1.87) Minor in possession 3.50 (2.08) First-cousin marriage 3.70 (2.08) Note. Arbitrariness ratings from Experiment 3. Mean ratings are followed in parentheses by standard deviations. Strict liability crimes are italicized.

Received November 15, 2015 Revision received July 11, 2016 Accepted July 16, 2016 Ⅲ This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.